Abstract

The paper presents and interprets the results of a well-powered one-time unconditional cash transfer experiment for low-income people in the US. The cash significantly increased recipient spending, but decreased survey measures of self-assessed wellbeing relative to control. Participant attrition limits the strength of the conclusions we can draw from this high-quality study. Overall, the study modestly shifted my views against one-time transfers and toward other policy tools for reducing poverty and inequality in wealthy societies.

Summary Measures

We asked evaluators to give some overall assessments, in addition to ratings across a range of criteria. See the evaluation summary “metrics” for a more detailed breakdown of this. See these ratings in the context of all Unjournal ratings, with some analysis, in our data presentation here.1

Overall assessment (See footnote2)

Journal rank tier, normative rating (0-5): On a ‘scale of journals’, what ‘quality of journal’ should this be published in?3 Note: 0= lowest/none, 5= highest/best.

Written report

The paper presents and interprets the results of a well-powered one-time unconditional cash transfer experiment in the US during 2020 and 2021. The experiment randomly assigned low-income people to receive either $2000, $500, or zero dollars, and then measured various outcomes pre- and post-treatment using participants’ bank data and a four-wave survey (with one wave being pre-treatment).

The results suggest that $2000 and $500 treatments each significantly increase recipient spending, especially on digital cash transfers and retail, with no effect on spending on ‘harmful’ items such as alcohol, tobacco, or gambling. However, the results also show significant decreases in the survey measures of financial, psychological, cognitive, and physical health in the cash groups relative to the control group.

The paper presents its main interpretation of the negative results on the survey measures using a model that shows that people who 1) have negatively biased expectations of the size of a negative financial shock and 2) are induced to monitor their finances more closely due to receiving a positive financial shock (e.g., the one-time cash transfer treatments in this study) might end up unhappier than they would have if they had not received the cash transfer. More concisely, the argument is that the “unhappy surprise” of learning more about negative shocks to one’s own finances can outweigh the “happy surprise” of receiving the one-time cash transfer.

Overall, the paper makes a strong contribution to the cash transfer literature for a number of reasons, including the study’s multiple treatment arms, its geographic context, its timing during the height of the COVID pandemic, and its measurement of a broad range of outcomes. The study also has some important limitations, including participant attrition and challenges in nailing down the mechanism driving the survey results.

Attrition

Sample attrition is the main limitation of the study. The Manski bounds for treatment effects are correspondingly large. As one example, the lower and upper possible bounds for the treatment effect of the $2000 transfer on participants’ financial health are -1.30 to 1.02, respectively. Lee (2009) bound coefficients are also wide, with 95% CIs often approximately [-0.35, 0.25]. These are, of course, very conservative bounding exercises, but they put the main estimates in helpful context.

While the bounding exercises give pause, the consistently significant negative estimates after multiple imputation (as proposed in the study’s pre-analysis plan) increase my confidence in the results a bit. Still, the fact that the treatment effect results are smaller after multiple imputation is consistent with attrition indeed biasing the paper’s main estimates in a negative direction.

The paper also offers helpful theorization of the mechanisms by which attrition might bias the results in a negative direction in Section C.4. Possibilities include things like treated and control individuals having different baseline conditions, or living in areas that are trending differently with respect to COVID or economic conditions. There remain many additional possible mechanisms based on unobservables, but the evidence presented in this section helps to allay concerns about some of the more straightforward possibilities for how attrition would negatively bias the results.

In summary, the paper does a nice job of dealing with attrition as best it can, but it remains the main obstacle to learning more about the true treatment effects. I don’t have any great insights to offer on this front, but I hope that more work is done in future studies to find ways of improving response rates in the field.

Social Welfare Effects - Self-Assessed Wellbeing

What does this study teach us about policy solutions to poverty and inequality? In this section I focus on broader interpretations related to self-assessed wellbeing. This study’s results suggest that one-time unconditional cash transfers are not effective at improving self-assessed quality of life among poor people within 15 weeks (it’s possible, though I would not say likely, that the one-time transfers positively affected the longer-term wellbeing of treated participants).

However, the social welfare effects of one-time transfers like these remain ambiguous. As I described earlier, the study argues that the non-positive effects on self-reported wellbeing stem from treated individuals being induced to more “actively” monitor their finances, which leads to “unhappy surprises” about negative shocks to one’s finances that outweigh the “happy surprise” of the one-time cash transfer. Readers will vary in how normatively good or bad they think this story is.

Another ambiguity of the broader welfare effects involves network effects. Treated individuals often shared or spent the money with/on family and friends, but we don’t have measures of self-reported wellbeing of the friends and families. Section D.6 shows that the treatment didn’t affect respondents’ relationships, but only with people “outside the household.” If the paper’s main mechanism is correct, then it is plausible that individuals close to the treated individual get an indirect version of the “happy surprise” of a one-time cash transfer while avoiding the “unhappy surprise” of learning about unexpected negative financial shocks.

Taken together, we still have a lot to learn about the broader effects of one-time cash transfers on aggregate (self-understood) wellbeing.

Social Welfare Effects - Reducing Poverty & Inequality

Does this study change my view of one-time cash transfers as a policy solution to poverty and inequality in wealthy country contexts? It is worth noting that cash transfers mechanically make people less poor, at least for some amount of time. The study is limited by its measurement timeframe, but recipients spent significantly more over the period (often some months) for which the study has bank account data. These outcomes reflect important policy goals in and of themselves.

The policy importance of cash transfers matters even if the cash does not increase recipients’ net worth. For example, the new Bartik et al (2024)[1] study has a different treatment (monthly cash transfers for three years), but finds no effect of the transfers on net worth because recipients took on more debt (often car loans) to match their newly increased incomes—but if recipients were better positioned or freer to access credit after receiving cash transfers, this would represent a policy success for the many who have called for greater equality in credit provision across class, race, and gender. For more discussion of the normative implications of access to credit and debt, see the work of legal scholar Abbye Atkinson. (It is also worth noting here that we do not know much about the very long term effects, on the order of decades, lifetimes, or across generations, of cash transfers.)

Conclusion

The “How Effective Is (More) Money?” paper makes a contribution to our understanding of one-time cash transfers to the poor in wealthy society contexts. The experimental design was strong, with a well-powered sample and many important repeated outcome measures. The paper also does a really nice job squeezing every possible ounce of inference out of the available data. Respondent attrition and uncertainty about mechanisms are common in field experiments, and scholars can use this paper as a guide for how to gauge the size of potential bias in estimates and adjudicate between potential mechanisms driving unexpected results.

The paper did indeed shift my beliefs about the medium-term effects of moderately sized one-time cash transfers on self-assessed wellbeing downward a bit, from an expectation of positive effects toward an expectation of zero effects. Attrition does weaken the signal conveyed by the estimates. But there still is a signal. A true positive effect is still possible in this case, but this would require the respondents who dropped out of the survey to have pretty exceptionally distinct trends in self-assessed wellbeing.

In terms of informing policy, this study did not change my views very substantially on the role of cash transfers as a tool for improving society. As I described earlier in this review, even a true negative effect of cash transfers on self-assessed wellbeing could be normatively offset by the spending effects or if there exist positive network effects for recipients’ family and friends. Cash transfers, both by giving money to low income families or by financing them through taxation, can be a tool to reduce economic inequality and the negative externalities that come with it. Furthermore, current and proposed cash transfers seem to work well in the US. This study doesn’t present direct evidence against the efficacy of a child tax credit, which, in one new prominent national proposal in the US, would be a one-time $6000 transfer. A plethora of difference-in-differences analyses of the Earned Income Tax Credit, such as those by Hilary Hoynes and her collaborators, show positive effects on infant and child health, educational outcomes, future earnings, and even reductions in children’s long term chances of being incarcerated.

However, societies’ policy choice is not simply a binary yes or no to cash transfers. There are other policy tools to increase spending by the poor and reduce economic inequality, and their trade-offs relative to cash transfers should be compared. This study’s findings modestly shift my views away from unconditional one-time cash transfers toward other policy tools that would increase worker bargaining power, like sectoral bargaining, other welfare state tools that increase redistribution while also reducing rent extraction, like universal single payer health insurance, or industrial policy. Still, as global wealth continues to increase, I see an important role for unconditional cash transfers.

References

[1]Bartik, A., Rhodes, E., Broockman, D., Krause, P., Miller, S., & Vivalt, E. (2024). The Impact of Unconditional Cash Transfers on Consumption and Household Balance Sheets: Experimental Evidence from Two US States. National Bureau of Economic Research. https://doi.org/10.3386/w32784

Evaluator details

1. How long have you been in this field?

   • 12 years

2. How many proposals and papers have you evaluated?

   • Hundreds